X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Frecgrs.c;h=fecf2c5406d6e26faf7145ef61b62ab6b88ebe26;hb=b88909df16157ed1e7859bc3fad6b01520d4865e;hp=42886984a227e7954d1138a6c011522d1e85c190;hpb=7c7aeb7843c5bec48a8928f07f6498fc12b3eb72;p=idzebra-moved-to-github.git diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c index 4288698..fecf2c5 100644 --- a/recctrl/recgrs.c +++ b/recctrl/recgrs.c @@ -1,5 +1,5 @@ -/* $Id: recgrs.c,v 1.99 2005-01-17 22:32:16 adam Exp $ - Copyright (C) 1995-2005 +/* $Id: recgrs.c,v 1.111 2006-06-13 12:02:15 adam Exp $ + Copyright (C) 1995-2006 Index Data ApS This file is part of the Zebra server. @@ -87,7 +87,7 @@ static int sp_range(struct source_parser *sp, data1_node *n, RecWord *wrd) /* 2nd arg: start */ if (!sp_expr(sp, n, &tmp_w)) return 0; - start = atoi_n(tmp_w.string, tmp_w.length); + start = atoi_n(tmp_w.term_buf, tmp_w.term_len); if (sp->lookahead == ',') { @@ -96,22 +96,22 @@ static int sp_range(struct source_parser *sp, data1_node *n, RecWord *wrd) /* 3rd arg: length */ if (!sp_expr(sp, n, &tmp_w)) return 0; - len = atoi_n(tmp_w.string, tmp_w.length); + len = atoi_n(tmp_w.term_buf, tmp_w.term_len); } else - len = wrd->length; + len = wrd->term_len; /* ) */ if (sp->lookahead != ')') return 0; sp_lex(sp); - if (wrd->string && wrd->length) + if (wrd->term_buf && wrd->term_len) { - wrd->string += start; - wrd->length -= start; - if (wrd->length > len) - wrd->length = len; + wrd->term_buf += start; + wrd->term_len -= start; + if (wrd->term_len > len) + wrd->term_len = len; } return 1; } @@ -134,13 +134,13 @@ static int sp_first(struct source_parser *sp, data1_node *n, RecWord *wrd) if (!sp_expr(sp, n, &search_w)) return 0; - for (i = 0; ilength; i++) + for (i = 0; iterm_len; i++) { int j; - for (j = 0; jlength; j++) - if (wrd->string[i+j] != search_w.string[j]) + for (j = 0; jterm_len; j++) + if (wrd->term_buf[i+j] != search_w.term_buf[j]) break; - if (j == search_w.length) /* match ? */ + if (j == search_w.term_len) /* match ? */ { if (min_pos == -1 || i < min_pos) min_pos = i; @@ -154,8 +154,8 @@ static int sp_first(struct source_parser *sp, data1_node *n, RecWord *wrd) if (min_pos == -1) min_pos = 0; /* the default if not found */ sprintf(num_str, "%d", min_pos); - wrd->string = nmem_strdup(sp->nmem, num_str); - wrd->length = strlen(wrd->string); + wrd->term_buf = nmem_strdup(sp->nmem, num_str); + wrd->term_len = strlen(wrd->term_buf); return 1; } @@ -167,8 +167,8 @@ static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) { if (n->which == DATA1N_data) { - wrd->string = n->u.data.data; - wrd->length = n->u.data.len; + wrd->term_buf = n->u.data.data; + wrd->term_len = n->u.data.len; } sp_lex(sp); } @@ -176,8 +176,8 @@ static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) { if (n->which == DATA1N_tag) { - wrd->string = n->u.tag.tag; - wrd->length = strlen(n->u.tag.tag); + wrd->term_buf = n->u.tag.tag; + wrd->term_len = strlen(n->u.tag.tag); } sp_lex(sp); } @@ -192,18 +192,18 @@ static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) if (!sp_expr(sp, n, &tmp_w)) return 0; - wrd->string = ""; - wrd->length = 0; + wrd->term_buf = ""; + wrd->term_len = 0; if (n->which == DATA1N_tag) { data1_xattr *p = n->u.tag.attributes; - while (p && strlen(p->name) != tmp_w.length && - memcmp (p->name, tmp_w.string, tmp_w.length)) + while (p && strlen(p->name) != tmp_w.term_len && + memcmp (p->name, tmp_w.term_buf, tmp_w.term_len)) p = p->next; if (p) { - wrd->string = p->value; - wrd->length = strlen(p->value); + wrd->term_buf = p->value; + wrd->term_len = strlen(p->value); } } if (sp->lookahead != ')') @@ -220,22 +220,26 @@ static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) } else if (sp->len > 0 && isdigit(*(unsigned char *)sp->tok)) { - wrd->string = nmem_malloc(sp->nmem, sp->len); - memcpy(wrd->string, sp->tok, sp->len); - wrd->length = sp->len; + char *b; + wrd->term_len = sp->len; + b = nmem_malloc(sp->nmem, sp->len); + memcpy(b, sp->tok, sp->len); + wrd->term_buf = b; sp_lex(sp); } else if (sp->len > 2 && sp->tok[0] == '\'' && sp->tok[sp->len-1] == '\'') { - wrd->length = sp->len - 2; - wrd->string = nmem_malloc(sp->nmem, wrd->length); - memcpy(wrd->string, sp->tok+1, wrd->length); + char *b; + wrd->term_len = sp->len - 2; + b = nmem_malloc(sp->nmem, wrd->term_len); + memcpy(b, sp->tok+1, wrd->term_len); + wrd->term_buf = b; sp_lex(sp); } else { - wrd->string = ""; - wrd->length = 0; + wrd->term_buf = ""; + wrd->term_len = 0; sp_lex(sp); } return 1; @@ -285,7 +289,7 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) yaz_log(YLOG_WARN, " Only attributes (@) are supported in xelm xpath predicates"); yaz_log(YLOG_WARN, "predicate %s ignored", p->u.relation.name); - return (1); + return 1; } attname = p->u.relation.name + 1; res = 0; @@ -337,6 +341,37 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) } +static int dfa_match_first(struct DFA_state **dfaar, const char *text) +{ + struct DFA_state *s = dfaar[0]; /* start state */ + struct DFA_tran *t; + int i; + const char *p = text; + unsigned char c; + + for (c = *p++, t = s->trans, i = s->tran_no; --i >= 0; t++) + { + if (c >= t->ch[0] && c <= t->ch[1]) + { + while (i >= 0) + { + /* move to next state and return if we get a match */ + s = dfaar[t->to]; + if (s->rule_no) + return 1; + /* next char */ + if (!c) + return 0; + c = *p++; + for (t = s->trans, i = s->tran_no; --i >= 0; t++) + if (c >= t->ch[0] && c <= t->ch[1]) + break; + } + } + } + return 0; +} + /* *ostrich* New function, looking for xpath "element" definitions in abs, by @@ -361,84 +396,52 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n) data1_node *nn; #ifdef ENHANCED_XELM struct xpath_location_step *xp; - #endif - char *pexpr = xmalloc(strlen(tagpath)+2); + char *pexpr = xmalloc(strlen(tagpath)+5); int ok = 0; - sprintf (pexpr, "%s\n", tagpath); - yaz_log(YLOG_DEBUG,"Checking tagpath %s",tagpath); - while (xpe) + sprintf (pexpr, "/%s\n", tagpath); + for (; xpe; xpe = xpe->next) { - struct DFA_state **dfaar = xpe->dfa->states; - struct DFA_state *s=dfaar[0]; - struct DFA_tran *t; - const char *p; - int i; - unsigned char c; - int start_line = 1; - - c = *pexpr++; t = s->trans; i = s->tran_no; - if ((c >= t->ch[0] && c <= t->ch[1]) || (!t->ch[0])) { - p = pexpr; - do { - if ((s = dfaar[t->to])->rule_no && - (start_line || s->rule_nno)) { - ok = 1; - break; - } - for (t=s->trans, i=s->tran_no; --i >= 0; t++) { - if ((unsigned) *p >= t->ch[0] && (unsigned) *p <= t->ch[1]) - break; - } - p++; - } while (i >= 0); - } - if (ok) - yaz_log(YLOG_DEBUG," xpath match %s",xpe->xpath_expr); - else - yaz_log(YLOG_DEBUG," xpath no match %s",xpe->xpath_expr); + int i; + ok = dfa_match_first(xpe->dfa->states, pexpr); - pexpr--; if (ok) { #ifdef ENHANCED_XELM /* we have to check the perdicates up to the root node */ xp = xpe->xpath; /* find the first tag up in the node structure */ - nn = n; while (nn && nn->which != DATA1N_tag) { - nn = nn->parent; - } + for (nn = n; nn && nn->which != DATA1N_tag; nn = nn->parent) + ; /* go from inside out in the node structure, while going backwards trough xpath location steps ... */ - for (i=xpe->xpath_len - 1; i>0; i--) { - - yaz_log(YLOG_DEBUG,"Checking step %d: %s on tag %s", - i,xp[i].part,nn->u.tag.tag); + for (i = xpe->xpath_len - 1; i>0; i--) + { + yaz_log(YLOG_DEBUG, "Checking step %d: %s on tag %s", + i, xp[i].part, nn->u.tag.tag); - if (!d1_check_xpath_predicate(nn, xp[i].predicate)) { - yaz_log(YLOG_DEBUG," Predicates didn't match"); + if (!d1_check_xpath_predicate(nn, xp[i].predicate)) + { + yaz_log(YLOG_DEBUG, " Predicates didn't match"); ok = 0; break; } - if (nn->which == DATA1N_tag) { + if (nn->which == DATA1N_tag) nn = nn->parent; - } } #endif - if (ok) { + if (ok) break; - } } - xpe = xpe->next; } xfree(pexpr); if (ok) { - yaz_log(YLOG_DEBUG,"Got it"); + yaz_log(YLOG_DEBUG, "Got it"); return xpe->termlists; } else { return NULL; @@ -468,73 +471,77 @@ static void index_xpath_attr (char *tag_path, char *name, char *value, char *structure, struct recExtractCtrl *p, RecWord *wrd) { - wrd->attrSet = VAL_IDXPATH; - wrd->attrUse = 1; - wrd->reg_type = '0'; - wrd->string = tag_path; - wrd->length = strlen(tag_path); + wrd->index_name = ZEBRA_XPATH_ELM_BEGIN; + wrd->index_type = '0'; + wrd->term_buf = tag_path; + wrd->term_len = strlen(tag_path); (*p->tokenAdd)(wrd); if (value) { - wrd->attrUse = 1015; - wrd->reg_type = 'w'; - wrd->string = value; - wrd->length = strlen(value); + wrd->index_name = ZEBRA_XPATH_ATTR_CDATA; + wrd->index_type = 'w'; + wrd->term_buf = value; + wrd->term_len = strlen(value); (*p->tokenAdd)(wrd); } - - wrd->attrUse = 2; - wrd->reg_type = '0'; - wrd->string = tag_path; - wrd->length = strlen(tag_path); + wrd->index_name = ZEBRA_XPATH_ELM_END; + wrd->index_type = '0'; + wrd->term_buf = tag_path; + wrd->term_len = strlen(tag_path); (*p->tokenAdd)(wrd); } -static void index_xpath (struct source_parser *sp, data1_node *n, - struct recExtractCtrl *p, - int level, RecWord *wrd, int use) +static void mk_tag_path_full(char *tag_path_full, size_t max, data1_node *n) { - int i; - char tag_path_full[1024]; size_t flen = 0; data1_node *nn; + + /* we have to fetch the whole path to the data tag */ + for (nn = n; nn; nn = nn->parent) + { + if (nn->which == DATA1N_tag) + { + size_t tlen = strlen(nn->u.tag.tag); + if (tlen + flen > (max - 2)) + break; + memcpy (tag_path_full + flen, nn->u.tag.tag, tlen); + flen += tlen; + tag_path_full[flen++] = '/'; + } + else + if (nn->which == DATA1N_root) + break; + } + tag_path_full[flen] = 0; +} + + +static void index_xpath(struct source_parser *sp, data1_node *n, + struct recExtractCtrl *p, + int level, RecWord *wrd, + char *xpath_index, + int xpath_is_start + ) +{ + int i; + char tag_path_full[1024]; int termlist_only = 1; data1_termlist *tl; int xpdone = 0; - - yaz_log(YLOG_DEBUG, "index_xpath level=%d use=%d", level, use); if ((!n->root->u.root.absyn) || - (n->root->u.root.absyn->enable_xpath_indexing)) { + (n->root->u.root.absyn->xpath_indexing == DATA1_XPATH_INDEXING_ENABLE)) { termlist_only = 0; } switch (n->which) { case DATA1N_data: - wrd->string = n->u.data.data; - wrd->length = n->u.data.len; + wrd->term_buf = n->u.data.data; + wrd->term_len = n->u.data.len; xpdone = 0; - flen = 0; - - /* we have to fetch the whole path to the data tag */ - for (nn = n; nn; nn = nn->parent) - { - if (nn->which == DATA1N_tag) - { - size_t tlen = strlen(nn->u.tag.tag); - if (tlen + flen > (sizeof(tag_path_full)-2)) - break; - memcpy (tag_path_full + flen, nn->u.tag.tag, tlen); - flen += tlen; - tag_path_full[flen++] = '/'; - } - else - if (nn->which == DATA1N_root) - break; - } - - tag_path_full[flen] = 0; + + mk_tag_path_full(tag_path_full, sizeof(tag_path_full), n); /* If we have a matching termlist... */ if (n->root->u.root.absyn && @@ -544,23 +551,23 @@ static void index_xpath (struct source_parser *sp, data1_node *n, { /* need to copy recword because it may be changed */ RecWord wrd_tl; - wrd->reg_type = *tl->structure; - /* this is the ! case, so structure is for the xpath index */ + wrd->index_type = *tl->structure; memcpy (&wrd_tl, wrd, sizeof(*wrd)); if (tl->source) sp_parse(sp, n, &wrd_tl, tl->source); - if (!tl->att) { - wrd_tl.attrSet = VAL_IDXPATH; - wrd_tl.attrUse = use; + if (!tl->index_name) + { + /* this is the ! case, so structure is for the xpath index */ + wrd_tl.index_name = xpath_index; if (p->flagShowRecords) { int i; printf("%*sXPath index", (level + 1) * 4, ""); printf (" XData:\""); - for (i = 0; i 40) + if (wrd_tl.term_len > 40) printf (" ..."); fputc ('\n', stdout); } @@ -569,22 +576,18 @@ static void index_xpath (struct source_parser *sp, data1_node *n, xpdone = 1; } else { /* this is just the old fashioned attribute based index */ - wrd_tl.attrSet = (int) (tl->att->parent->reference); - wrd_tl.attrUse = tl->att->locals->local; + wrd_tl.index_name = tl->index_name; if (p->flagShowRecords) { int i; printf("%*sIdx: [%s]", (level + 1) * 4, "", tl->structure); - printf("%s:%s [%d] %s", - tl->att->parent->name, - tl->att->name, tl->att->value, - tl->source); + printf("%s %s", tl->index_name, tl->source); printf (" XData:\""); - for (i = 0; i 40) + if (wrd_tl.term_len > 40) printf (" ..."); fputc ('\n', stdout); } @@ -597,40 +600,23 @@ static void index_xpath (struct source_parser *sp, data1_node *n, or no ! in the termlist, and default indexing is enabled... */ if (!p->flagShowRecords && !xpdone && !termlist_only) { - wrd->attrSet = VAL_IDXPATH; - wrd->attrUse = use; - wrd->reg_type = 'w'; + wrd->index_name = xpath_index; + wrd->index_type = 'w'; (*p->tokenAdd)(wrd); } break; case DATA1N_tag: - flen = 0; - for (nn = n; nn; nn = nn->parent) - { - if (nn->which == DATA1N_tag) - { - size_t tlen = strlen(nn->u.tag.tag); - if (tlen + flen > (sizeof(tag_path_full)-2)) - break; - memcpy (tag_path_full + flen, nn->u.tag.tag, tlen); - flen += tlen; - tag_path_full[flen++] = '/'; - } - else if (nn->which == DATA1N_root) - break; - } - + mk_tag_path_full(tag_path_full, sizeof(tag_path_full), n); - wrd->reg_type = '0'; - wrd->string = tag_path_full; - wrd->length = flen; - wrd->attrSet = VAL_IDXPATH; - wrd->attrUse = use; + wrd->index_type = '0'; + wrd->term_buf = tag_path_full; + wrd->term_len = strlen(tag_path_full); + wrd->index_name = xpath_index; if (p->flagShowRecords) { printf("%*s tag=", (level + 1) * 4, ""); - for (i = 0; ilength && i < 40; i++) - fputc (wrd->string[i], stdout); + for (i = 0; iterm_len && i < 40; i++) + fputc (wrd->term_buf[i], stdout); if (i == 40) printf (" .."); printf("\n"); @@ -641,20 +627,26 @@ static void index_xpath (struct source_parser *sp, data1_node *n, data1_termlist *tl; int do_xpindex; - tag_path_full[flen] = 0; - - /* Add tag start/end xpath index, only when there is a ! in the apropriate xelm - directive, or default xpath indexing is enabled */ - if (!(do_xpindex = 1 - termlist_only)) { - if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) { - for (; tl; tl = tl->next) { if (!tl->att) {do_xpindex = 1;} } + /* Add tag start/end xpath index, only when there is a ! in + the apropriate xelm directive, or default xpath indexing + is enabled + */ + if (!(do_xpindex = 1 - termlist_only)) + { + if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) + { + for (; tl; tl = tl->next) + { + if (!tl->index_name) + do_xpindex = 1; + } } } if (do_xpindex) { (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ } - if (use == 1) /* only for the starting tag... */ + if (xpath_is_start == 1) /* only for the starting tag... */ { #define MAX_ATTR_COUNT 50 data1_termlist *tll[MAX_ATTR_COUNT]; @@ -672,11 +664,10 @@ static void index_xpath (struct source_parser *sp, data1_node *n, int do_xpindex = 1 - termlist_only; data1_termlist *tl; char attr_tag_path_full[1024]; - int int_len = flen; /* this could be cached as well */ - sprintf (attr_tag_path_full, "@%s/%.*s", - xp->name, int_len, tag_path_full); + sprintf (attr_tag_path_full, "@%s/%s", + xp->name, tag_path_full); tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n); @@ -685,19 +676,19 @@ static void index_xpath (struct source_parser *sp, data1_node *n, if ((tl = tll[i])) { for (; tl; tl = tl->next) - { - if (!tl->att) - do_xpindex = 1; + { + if (!tl->index_name) + do_xpindex = 1; } } if (do_xpindex) { /* attribute (no value) */ - wrd->reg_type = '0'; - wrd->attrUse = 3; - wrd->string = xp->name; - wrd->length = strlen(xp->name); + wrd->index_type = '0'; + wrd->index_name = ZEBRA_XPATH_ATTR_NAME; + wrd->term_buf = xp->name; + wrd->term_len = strlen(xp->name); wrd->seqno--; (*p->tokenAdd)(wrd); @@ -709,11 +700,11 @@ static void index_xpath (struct source_parser *sp, data1_node *n, strcpy (comb, xp->name); strcat (comb, "="); strcat (comb, xp->value); - - wrd->attrUse = 3; - wrd->reg_type = '0'; - wrd->string = comb; - wrd->length = strlen(comb); + + wrd->index_name = ZEBRA_XPATH_ATTR_NAME; + wrd->index_type = '0'; + wrd->term_buf = comb; + wrd->term_len = strlen(comb); wrd->seqno--; (*p->tokenAdd)(wrd); @@ -726,32 +717,31 @@ static void index_xpath (struct source_parser *sp, data1_node *n, for (xp = n->u.tag.attributes; xp; xp = xp->next) { data1_termlist *tl; char attr_tag_path_full[1024]; - int int_len = flen; int xpdone = 0; - sprintf (attr_tag_path_full, "@%s/%.*s", - xp->name, int_len, tag_path_full); + sprintf (attr_tag_path_full, "@%s/%s", + xp->name, tag_path_full); if ((tl = tll[i])) { /* If there is a termlist given (=xelm directive) */ for (; tl; tl = tl->next) { - if (!tl->att) { + if (!tl->index_name) + { /* add xpath index for the attribute */ index_xpath_attr (attr_tag_path_full, xp->name, xp->value, tl->structure, p, wrd); xpdone = 1; } else { - /* add attribute based index for the attribute */ - if (xp->value) { - wrd->attrSet = (int) - (tl->att->parent->reference); - wrd->attrUse = tl->att->locals->local; - wrd->reg_type = *tl->structure; - wrd->string = xp->value; - wrd->length = strlen(xp->value); + /* index attribute value (only path/@attr) */ + if (xp->value) + { + wrd->index_name = tl->index_name; + wrd->index_type = *tl->structure; + wrd->term_buf = xp->value; + wrd->term_len = strlen(xp->value); (*p->tokenAdd)(wrd); } } @@ -796,34 +786,30 @@ static void index_termlist (struct source_parser *sp, data1_node *par, for (; tlist; tlist = tlist->next) { /* consider source */ - wrd->string = 0; + wrd->term_buf = 0; assert(tlist->source); sp_parse(sp, n, wrd, tlist->source); - if (wrd->string && wrd->length) + if (wrd->term_buf && wrd->term_len) { if (p->flagShowRecords) { int i; printf("%*sIdx: [%s]", (level + 1) * 4, "", tlist->structure); - printf("%s:%s [%d] %s", - tlist->att->parent->name, - tlist->att->name, tlist->att->value, - tlist->source); + printf("%s %s", tlist->index_name, tlist->source); printf (" XData:\""); - for (i = 0; ilength && i < 40; i++) - fputc (wrd->string[i], stdout); + for (i = 0; iterm_len && i < 40; i++) + fputc (wrd->term_buf[i], stdout); fputc ('"', stdout); - if (wrd->length > 40) + if (wrd->term_len > 40) printf (" ..."); fputc ('\n', stdout); } else { - wrd->reg_type = *tlist->structure; - wrd->attrSet = (int) (tlist->att->parent->reference); - wrd->attrUse = tlist->att->locals->local; + wrd->index_type = *tlist->structure; + wrd->index_name = tlist->index_name; (*p->tokenAdd)(wrd); } } @@ -878,7 +864,8 @@ static int dumpkeys_r(struct source_parser *sp, index_termlist(sp, n, n, p, level, wrd); /* index start tag */ if (n->root->u.root.absyn) - index_xpath(sp, n, p, level, wrd, 1); + index_xpath(sp, n, p, level, wrd, ZEBRA_XPATH_ELM_BEGIN, + 1 /* is start */); } if (n->child) @@ -906,13 +893,15 @@ static int dumpkeys_r(struct source_parser *sp, if (par) index_termlist(sp, par, n, p, level, wrd); - index_xpath(sp, n, p, level, wrd, 1016); + index_xpath(sp, n, p, level, wrd, ZEBRA_XPATH_CDATA, + 0 /* is start */); } if (n->which == DATA1N_tag) { /* index end tag */ - index_xpath(sp, n, p, level, wrd, 2); + index_xpath(sp, n, p, level, wrd, ZEBRA_XPATH_ELM_END, + 0 /* is start */); } if (p->flagShowRecords && n->which == DATA1N_root) @@ -1451,3 +1440,11 @@ int zebra_grs_retrieve(void *clientData, struct recRetrieveCtrl *p, return 0; } +/* + * Local variables: + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ +